BB0562 is a nutritional virulence determinant with lipase activity important for Borrelia burgdorferi infection and survival in fatty acid deficient environments

The Lyme disease spirochete Borrelia burgdorferi relies on uptake of essential nutrients from its host environments for survival and infection. Therefore, nutrient acquisition mechanisms constitute key virulence properties of the pathogen, yet these mechanisms remain largely unknown. In vivo expression technology applied to B. burgdorferi (BbIVET) during mammalian infection identified gene bb0562, which encodes a hypothetical protein comprised of a conserved domain of unknown function, DUF3996. DUF3996 is also found across adjacent encoded hypothetical proteins BB0563 and BB0564, suggesting the possibility that the three proteins could be functionally related. Deletion of bb0562, bb0563 and bb0564 individually and together demonstrated that bb0562 alone was important for optimal disseminated infection in immunocompetent and immunocompromised mice by needle inoculation and tick bite transmission. Moreover, bb0562 promoted spirochete survival during the blood dissemination phase of infection. Gene bb0562 was also found to be important for spirochete growth in low serum media and the growth defect of Δbb0562 B. burgdorferi was rescued with the addition of various long chain fatty acids, particularly oleic acid. In mammals, fatty acids are primarily stored in fat droplets in the form of triglycerides. Strikingly, addition of glyceryl trioleate, the triglyceride form of oleic acid, to the low serum media did not rescue the growth defect of the mutant, suggesting bb0562 may be important for the release of fatty acids from triglycerides. Therefore, we searched for and identified two canonical GXSXG lipase motifs within BB0562, despite the lack of homology to known bacterial lipases. Purified BB0562 demonstrated lipolytic activity dependent on the catalytic serine residues within the two motifs. In sum, we have established that bb0562 is a novel nutritional virulence determinant, encoding a lipase that contributes to fatty acid scavenge for spirochete survival in environments deficient in free fatty acids including the mammalian host.

Introduction Lyme disease, caused by the spirochete Borrelia (Borreliella) burgdorferi, is the leading vectorborne bacterial illness in the world [1]. B. burgdorferi is naturally maintained in an enzootic cycle between an arthropod vector, Ixodes scapularis ticks, and diverse small vertebrate reservoir hosts [2]. Humans become infected with B. burgdorferi through the bite of an infected tick [3]. If left untreated, B. burgdorferi disseminates from the bite site, transiently through the blood, to distal tissues leading to the debilitating clinical manifestations of Lyme disease such as arthritis, carditis, facial nerve palsy and encephalopathy [4].
For B. burgdorferi to achieve a disseminated infection the pathogen must adapt to physical conditions of the host environment, scavenge critical nutrients, overcome host innate and adaptive immune defenses, and interact with key host cell molecules to effectively move through and colonize distal tissues [5,6]. Broad research efforts in the field have resulted in identification of numerous B. burgdorferi proteins important for spirochete dissemination [5,7]. Largely, these studies have been guided by hypotheses based on protein sequence identity and conserved domain analyses. However, B. burgdorferi harbors a reduced genome lacking many canonical metabolic, virulence, and host defense evasion proteins. Furthermore, a large number of B. burgdorferi genes encode hypothetical proteins of unknown function which are uniquely conserved among Borrelia species [8,9]. Together these observations suggest that B. burgdorferi has evolved novel mechanisms to survive in the host.
As a result of its condensed genome, B. burgdorferi has significantly reduced metabolic and biosynthetic capabilities [9,10] and therefore is an obligate pathogen, dependent on its tick vector and mammalian host for salvage of essential nutrients. The competition between pathogens and the metabolism of their hosts for limited nutrients in the host environment has been coined "nutritional virulence" [11]. There is a growing body of evidence that nutrient acquisition functions are key virulence determinants for B. burgdorferi [12]. Uptake systems for glucose [13], oligopeptides [14], purines [15], manganese [16] and riboflavin [17] are essential for mammalian infection. Manganese uptake is also required for B. burgdorferi survival in ticks [16]. In addition, glycerol utilization is critical for spirochete persistence in ticks [18]. B. burgdorferi lacks genes for the de novo biosynthesis of fatty acids and cholesterol [9,10], which are critical components of the spirochete membrane, lipoproteins, glycolipids and lipid rafts [19][20][21][22][23][24]. B. burgdorferi has been shown to acquire cholesterol from eukaryotic cells through direct contact [21], yet the genetic basis for this process remains unknown. Gene bb0646 encodes a lipase with specificity for saturated and polyunsaturated substrates, as well as has hemolytic activity [25], suggesting a possible role in fatty acid acquisition. The lipolytic activity of BB0646 contributes to optimal infectivity of B. burgdorferi [25], supporting fatty acid scavenge as a virulence property of the spirochete.
In order to elucidate novel factors important for B. burgdorferi survival and infection we previously sought to identify transcripts expressed by the pathogen during an active mammalian infection through application of in vivo expression technology (BbIVET) [26,27]. This in vivo genetic screen identified 233 infection-active putative promoter sequences, referred to as B. burgdorferi in vivo expressed sequences (Bbives). Ninety-one Bbives mapped within 300 nts upstream of annotated start codons for open reading frames (ORFs) [27,28], suggesting that these sequences serve as promoters for the proximal genes during infection. Approximately 33% of the 91 Bbives mapped upstream of genes annotated to encode hypothetical proteins of unknown function. Many of these genes have yet to be investigated for their contributions to B. burgdorferi infection.
Herein, we targeted the BbIVET-identified gene of unknown function, bb0562 and putatively related adjacent genes bb0563 and bb0564 for deletion and examination throughout the B. burgdorferi enzootic cycle. We demonstrated that of these three genes, bb0562 alone was important for disseminated infection in mice. We established that bb0562 is a novel nutritional virulence determinant, encoding a lipase that contributes to fatty acid scavenge for spirochete survival in fatty acid deficient environments including the mammalian host.

BbIVET-identified gene bb0562 is constitutively expressed throughout the B. burgdorferi enzootic cycle and encodes an outer membrane-associated protein
In vivo expression technology for B. burgdorferi (BbIVET) identified an infection-active putative promoter sequence immediately upstream of gene bb0562 [27,28]. Furthermore, using global 5 0 end RNA-seq analysis we detected a primary transcription start site (pTSS) for bb0562 located at the 3 0 end of the BbIVET-identified sequence (Fig 1A) [28]. Gene bb0562 is annotated to encode a 180 amino acid (aa) hypothetical protein of unknown function and to be comprised of the conserved domain of unknown function DUF3996 (pfam13161) [29]. DUF3996 is also found across adjacent encoded proteins BB0563 (172 aa) and BB0564 (201 aa), raising the possibility of shared function between the three proteins. BB0562, BB0563 and BB0564 each demonstrate high amino acid identity within Borrelia burgdorferi sensu stricto isolates (97-99%) and across the Borrelia burgdorferi sensu lato complex (78-97%). Despite

PLOS PATHOGENS
bb0562 encodes a lipase important for Borrelia infection Gene bb0562 is constitutively expressed throughout the enzootic cycle and produces an outer membrane-associated, non-surface exposed protein. (A) RNA sequencing browser image of the bb0562-bb0564 locus displaying sequencing reads and identifying primary transcription start sites (pTSS) from 5 0 RNA-seq [28]. Tracks represent an overlay of two biological replicates. Read count ranges are shown in the top of each frame. The chromosome nucleotide coordinates, relative orientation of annotated ORFs (wide black arrows), Bbive sequence (wide blue arrow) [27,28], and pTSSs (bent black arrows) [28] are indicated. (B) Northern analysis of the bb0562, bb0563, and bb0564 transcripts. Total RNA from wild-type B. burgdorferi was separated on an agarose gel, transferred to a membrane, and probed for the indicated RNAs (RNAs were probed sequentially on the same membrane). Size markers are displayed on the left, which is aligned for all blots. White asterisks in panel A indicate the approximate locations of the oligonucleotide probes. (C) Reverse transcription (RT)-qPCR analysis of bb0562, bb0563 and bb0564 expression in infected tick and mouse bladder tissue samples. Genes flaB, ospA and ospC served as controls [62]. Copy numbers for each gene target were normalized to recA mRNA copies. Data are presented as the average of biological triplicate samples ± standard deviation. Expression levels across samples for each target gene were compared by one-way ANOVA with Tukey's multiple comparison test, GraphPad Prism 9.0.0. (n.s., not significant). (D) Total protein lysate (T) of wild-type (WT) or B. burgdorferi producing BB0562-3XFLAG (bb0562-3Xflag) was fractionated into the soluble (S) and membrane (M) components using ultracentrifugation. Protein fractions from equivalent numbers of spirochetes were subjected to SDS-PAGE and analyzed by immunoblot with antibodies against BB0562-3XFLAG (αFLAG), membrane-associated FlaB (αFlaB), or cytoplasmic SodA (αSodA). Molecular weights are shown in kilodaltons (kDa). (E) Total protein lysate (T) of wild-type (WT) or B. burgdorferi producing BB0562-containing the same conserved domain, which spans the majority of all three proteins, the proteins share limited amino acid identity with each other (24-29%). Genes bb0562 and bb0563 are separated by 164 bps and genes bb0563 and bb0564 by 82 bps. We previously identified a pTSS for bb0563 but not bb0564 [28] (Fig 1A). Northern blot analysis revealed primarily monocistronic transcripts for each of the three genes ( Fig 1B). Furthermore, genes bb0562, bb0563 and bb0564 were found to be constitutively expressed across all stages of the tickmouse enzootic cycle, though the overall levels of bb0562 were lower than that of bb0563 and bb0564 (Fig 1C).
Bioinformatics analysis suggests that BB0562 may harbor multiple transmembrane domains [30][31][32]. Furthermore, proteomic analysis of spirochete membranes has suggested that BB0562 is membrane associated in B. burgdorferi (strain B31) [33], B. afzelii (strain K78), and B. garinii (strain PBi) [34]. To experimentally test BB0562 protein production and localization in the spirochete, the sequence for a triple flag epitope (3Xflag) was added to the C-terminus of bb0562. Western blot analysis of total protein lysate from the bb0562-3Xflag clone using anti-FLAG antibodies specifically detected a~20 kDa protein, the predicted molecular mass of BB0562 ( Fig 1D). BB0562-3XFLAG, like membrane-associated FlaB but unlike soluble SodA, was found to be restricted to the spirochete membrane fraction (Fig 1D). To further dissect BB0562 localization, outer membrane and protoplasmic cylinder fractions were collected. Similar to outer membrane protein P66, but in contrast to inner membrane protein OppAIV, BB0562-3XFLAG was detected in both fractions, suggesting that BB0562 is associated with the outer membrane ( Fig 1E). Proteinase K treatment of intact bb0562-3Xflag spirochetes did not affect the amount or size of BB0562-3XFLAG (Fig 1F). Similarly, the periplasmic FlaB protein was unaffected by proteinase K treatment, whereas this treatment resulted in complete loss of the known surface exposed outer membrane protein VlsE. In contrast, BB0562-3XFLAG was susceptible to proteolysis upon permeabilization of the spirochetes with SDS ( Fig 1F). Together these data support a model in which BB0562 is associated with the outer membrane or a protein(s) on the inner side of the outer membrane and is likely not a transmembrane outer membrane protein or surface exposed.

Gene bb0562 provides a physiological function important for B. burgdorferi disseminated infection
In order to assess the contributions of bb0562, bb0563 and bb0564 to B. burgdorferi biology, targeted deletion of each gene individually or all three together was carried out by allelic exchange and each deletion was genetically complemented by restoration of a wild-type copy of the gene(s) in the mutant locus (S1A Fig). The mutant and complement clones were verified by end-point PCR (S1B Fig) and RT-qPCR analysis (S2 Fig). For the most part, expression levels of adjacent genes were unaltered in the mutant clones. Yet, expression of bb0563 and bb0564 tended to be reduced in the absence of bb0562 (S2 Fig). In the complement clones, wild-type or near wild-type levels of expression were restored for each of the target genes, with the exception of the amounts of expression of bb0562 and bb0564 in bb0562-bb0564+ B. burgdorferi, which were significantly reduced compared to that of wild-type (S2 Fig). No defect was detected for any of the mutants during growth in complete BSKII liquid medium (S3 Fig) and 3XFLAG (bb0562-3Xflag) was fractionated into the outer membrane (OM) and protoplasmic cylinder (PC) components by discontinuous sucrose gradient. Protein fractions from equivalent numbers of spirochetes were subjected to SDS-PAGE and analyzed by immunoblot with antibodies against BB0562-3XFLAG (αFLAG), OMassociated P66 (αP66), or PC-associated OppAIV (αOppAIV). Molecular weights are shown in kilodaltons (kDa). (F) Equal numbers of B. burgdorferi producing BB0562-3XFLAG were either treated with (+) or without (-) 200 μg/ml proteinase K (PK) in the presence (+) or absence (-) of 0.1% SDS. Samples were subjected to SDS-PAGE and analyzed by immunoblot with antibodies against BB0562-3XFLAG (αFLAG), periplasmic-associated FlaB (αFlaB), or outer surface-associated VlsE (αVlsE). Protein standards in kilodaltons (kDa) are indicated. https://doi.org/10.1371/journal.ppat.1009869.g001

PLOS PATHOGENS
bb0562 encodes a lipase important for Borrelia infection no obvious morphological alterations or motility defects were observed under routine dark field microscopy. Groups of mice were needle inoculated with 10 4 of each of the B. burgdorferi mutant and complement clones and assessed for infection after 3 weeks by reisolation of spirochetes from tissues. B. burgdorferi clones lacking bb0562 (Δbb0562 and Δbb0562-bb0564) demonstrated attenuated disseminated infection with an average of 49% and 20% of tissues positive for spirochete reisolation, respectively, compared to 96-100% for the B. burgdorferi clones carrying bb0562 (Fig 2A and S2 Table). Genes bb0563 and bb0564 alone were found to be dispensable for mouse infection. Although these data suggested that the triple mutant may be more highly attenuated than the single Δbb0562 mutant, the infectivity defect of Δbb0562-bb0564 B. burgdorferi was fully restored with the reintroduction of bb0562 alone in trans on the shuttle vector pBSV2G (Δbb0562-bb0564/bb0562+) as assessed by reisolation of spirochetes from tissues ( Fig 2B) and quantification of spirochete loads in tissues by quantitative PCR (S4 Fig), indicating that of genes bb0562-bb0564, bb0562 was sufficient for optimal infection.
Productive disseminated infection relies on successful spirochete population expansion in the skin site of infection and hematogenous spread to the tissues [35]. To evaluate the contributions of gene bb0562 to early steps in dissemination, spirochete burdens were quantified in the blood and skin site of inoculation at days 6 and 7 post-inoculation, respectively. Strikingly, the numbers of Δbb0562 and Δbb0562-bb0564 B. burgdorferi in the blood were significantly reduced compared to that of wild-type B. burgdorferi ( Fig 2C). In contrast, there was no bb0562-dependent defect observed for B. burgdorferi loads in the skin inoculation site ( Fig  2D). These data suggested that bb0562 was dispensable for localized expansion of B. burgdorferi in the skin but important for spirochete survival and dissemination in the blood.
To gain further insight into the potential barriers to Δbb0562 B. burgdorferi infectivity, we examined the ability of the mutant to infect immunodeficient mice. NOD-scid IL2Rgamma null (NSG) mice lack key components of innate immunity and the entire adaptive immune system [36]. The Δbb0562 B. burgdorferi mutant was defective for infection of NSG mice and resulted in only 33% of tissues positive for spirochete reisolation (Fig 2E and S2 Table). All tissues from NSG mice inoculated with bb0562 + B. burgdorferi were positive for spirochete reisolation ( Fig  2E and S2 Table). These data indicated that the immunocompromised state of the NSG mice was not sufficient to rescue the infectivity defect of Δbb0562 spirochetes and suggested that the molecular function of BB0562 was likely to be metabolic in nature rather than a mechanism of immune evasion.

Gene bb0562 contributes to B. burgdorferi murine infection by tick bite
Due to the attenuation of Δbb0562 B. burgdorferi for infection of both immunocompetent and immunocompromised mice by needle inoculation, we were interested in determining the contribution of bb0562 to B. burgdorferi survival throughout the tick-mouse infectious cycle. Naïve Ixodes scapularis larvae were artificially infected with the B. burgdorferi clones. All clones colonized the unfed larvae with the same efficiency ( Fig 3A). Moreover, no defect in spirochete load was detected at any tick life stage for any of the mutants (Fig 3B and 3D). Compared to wild-type B. burgdorferi, a significant increase in the number of spirochetes was detected for Δbb0563 B. burgdorferi and Δbb0564 B. burgdorferi in replete larvae and/or unfed nymphs, respectively. Yet, there were no statistical differences in spirochete number in fed nymphs for any of the mutants compared to that of wild-type B. burgdorferi. Overall, these data suggested that genes bb0562-bb0564 were dispensable for B. burgdorferi survival and replication in ticks. However, consistent with the needle inoculation studies, bb0562 was found to be important for mouse infection by tick bite. Mice fed on by Δbb0562-infected nymphs demonstrated low infection rates, as determined by spirochete reisolation from tissues, compared to mice fed on PLOS PATHOGENS bb0562 encodes a lipase important for Borrelia infection  Table). These data further supported a significant role for bb0562 in B. burgdorferi mammalian infectivity.

bb0562 is critical for B. burgdorferi growth in fatty acid limited conditions
Although we did not detect a growth phenotype for any of the mutants in complete BSKII liquid medium (S3 and S5A Figs), over the course of our studies we observed a variable defect in the growth of Δbb0562 and Δbb0562-bb0564 B. burgdorferi in complete solid BSK medium. To explore this observation further, aliquots of spirochetes collected at different time points across a 166-hour growth curve (S5A Fig) were quantified for colony forming units in both solid medium and liquid medium in a 96-well plate format. The data were reported as the percent survival in solid medium, which represented the number of colonies in solid medium normalized to the number of colonies by limiting dilution in complete liquid medium in 96-well plates (see Materials and Methods for more details). At mid-log phase (48-hour time point) Δbb0562 and Δbb0562-bb0564 B. burgdorferi both demonstrated significant survival defects in solid medium relative to wild-type B. burgdorferi, whereas the mutant clones harboring bb0562 did not ( Fig 4A). The defects of the bb0562 mutant clones were the most pronounced for spirochetes in mid-log and early stationary phase (S5 Fig). Recently, bacterial colony formation in solid media has been linked to fatty acid availability [37]. Because the B. burgdorferi genome lacks the genes for de novo synthesis of fatty acids [9], we hypothesized that bb0562 may contribute to B. burgdorferi's ability to scavenge fatty acids from the growth environment. Furthermore, we reasoned that rabbit serum likely provides the major source of fatty acids in Borrelia media and the concentration of fatty acids in rabbit serum could vary from lot to lot leading to the variation in the growth defect that we observed. Consistent with this notion, reduction in the amount of rabbit serum in solid BSK medium from 4% (complete) to 1.8% (low serum) resulted in a decreased percent survival of the Δbb0562 mutant relative to the wild-type clone in the same condition ( Fig 4B). Interestingly, the bb0562-dependent growth defect in low serum was not limited to growth in solid medium. This phenotype was also detected for Δbb0562 and Δbb0562-bb0564 B. burgdorferi grown in liquid medium containing only 1.8% serum ( Fig 4C). Similar to the mouse infection defect of Δbb0562-bb0564 B. burgdorferi, reintroduction of bb0562 alone to the triple mutant was sufficient to restore wild-type levels of growth in low serum medium ( Fig 4C). The long chain monounsaturated fatty acid, oleic acid (c18:1n9), long chain polyunsaturated fatty acid, linoleic acid (c18:2n6), and the saturated fatty acid, palmitic acid (c16:0), have been shown to comprise the vast majority of the fatty acid composition of B. burgdorferi grown in BSK media [38]. Each of these three fatty acids, or cholesterol, a significant non-fatty acid sterol lipid present in the B. burgdorferi membrane [21], was added to the low serum medium and the B. burgdorferi clones were assessed for growth. Strikingly, addition of oleic acid fully restored the bb0562-dependent growth defect in both low serum solid and liquid media to that of wild-type (Fig 4D and 4E). Similarly, a partial rescue of Δbb0562 B.

Fig 3. Genes bb0562-bb0564 are dispensable for B. burgdorferi survival and replication in ticks but bb0562 is important for mouse infection by tick bite transmission.
Naïve Ixodes scapularis larval ticks were artificially infected with B. burgdorferi clones by immersion. B. burgdorferi densities in ticks were assessed at each stage of tick development before and after feeding to repletion on mice by performing quantitative PCR analysis of total DNA isolated from ticks (A-C) or plating of dilutions of whole triturated ticks in solid or liquid BSK medium (D). Statistical analyses were performed using the one-way ANOVA with Dunnett's multiple comparisons test to WT, GraphPad Prism 9.0.0. ( �� p<0.01). (E) Cohorts of 25 B. burgdorferi infected nymphs were fed to repletion on groups of 3-7 naïve C3H/HeN mice per B. burgdorferi clone. Mice were assessed for infection three weeks post-tick feeding. Ear, heart, bladder, and joint tissues were collected and analyzed for spirochete reisolation in culture medium. Numbers above each data column represent the total number of culture positive tissues out of the total number of tissues analyzed. https://doi.org/10.1371/journal.ppat.1009869.g003

PLOS PATHOGENS
bb0562 encodes a lipase important for Borrelia infection burgdorferi growth in low serum solid medium occurred with the addition of linoleic acid or palmitic acid, but not with the addition of cholesterol (Fig 4D), together supporting the fatty acid dependence of the phenotype. In mammals, fatty acids are primarily stored in fat droplets in the form of triglycerides [39]. To model this, we examined the ability of the B. burgdorferi clones lacking bb0562 to grow in low serum liquid medium supplemented with the triglyceride glyceryl trioleate, which is comprised of three oleic acid molecules and a glycerol core. Unlike free oleic acid, glyceryl trioleate had no effect on the growth defect of Δbb0562 B. burgdorferi (Fig 4E), suggesting that free fatty acids are required to support the growth of the mutant and that bb0562 may be important for release of fatty acids from triglycerides for spirochete growth.

BB0562 is a lipolytic enzyme
Triacylglycerol lipases target hydrolysis of triacylglycerol to glycerol and free fatty acid [40] and typically have an α/β hydrolase fold and contain the active-site consensus motif GXSXG or GDSL [41,42]. Although the BB0562 protein lacks both overall primary sequence and tertiary structural homology to known bacterial lipases, visual inspection of the BB0562 amino acid sequence revealed two GXSXG motifs, with putative catalytic serine residues at amino acids 34 and 58 (Fig 5A and 5B). No lipase motifs were found in the amino acid sequences of BB0563 or BB0564. Recombinant BB0562 was produced in E. coli and purified as an N-terminal fusion to the highly soluble Nus protein along with N-and C-terminal 6XHis tags (S6 Fig). In addition, we generated mutant BB0562 proteins in which the putative catalytic serine residues alone or together were changed to alanine(s) or threonine(s) (Fig 5B). All seven recombinant proteins, along with the Nus protein alone, were examined for lipolytic activity in a fluorometric lipase assay. Strikingly, wild-type BB0562 demonstrated lipase activity significantly greater than that of the serine mutants and Nus alone ( Fig 5C). These data indicate that BB0562 is a lipolytic enzyme whose activity is dependent on the serine residues in two GXSXG catalytic motifs.

Discussion
B. burgdorferi is an obligate pathogen, dependent on its host environments for essential nutrients for survival. Therefore, nutrient scavenging mechanisms are critical to B. burgdorferi

PLOS PATHOGENS
bb0562 encodes a lipase important for Borrelia infection infection and pathogenesis. BbIVET identified numerous B. burgdorferi genes of unknown function that are expressed during mouse infection, including bb0562, which we have demonstrated encodes a novel lipase that promotes B. burgdorferi survival in environments low in free fatty acids, including the mammalian host. BB0562 adds to the arsenal of nutritional virulence determinants of B. burgdorferi important for the pathogen to be maintained in the tickmouse enzootic cycle and to cause disseminated disease.
Lipolytic enzymes are broadly present across pathogenic and nonpathogenic bacterial species [40,42]. A number of pathogens rely on lipases to hydrolyze host lipids for release of fatty acids for nutrients and several of these enzymes have also been shown to be important for infection. For example, Mycobacterium tuberculosis virulence, persistence, and drug-resistance have been linked to lipid acquisition and metabolism [43]. M. tuberculosis is predicted to possess up to 30 lipolytic enzymes [44,45], some of which contribute to virulence mechanisms such as intracellular growth, bacterial loads in mice and recovery from dormancy [46][47][48][49][50]. Pseudomonas aeruginosa lipases, LipA and LipC, influence motility and biofilm formation as well as the production of virulence factors pyoverdine and rhamnolipid [51,52]. Similarly, staphylococcal lipases play roles in biofilm formation and host cell invasion [53]. Moreover, phospholipases of numerous pathogenic bacteria, which target phospholipids in eukaryotic membranes, have been associated with cytotoxicity, host cell invasion, tissue colonization, hemolytic activity, phagosome escape and necrosis [54]. In total bacterial lipases are key components of host-pathogen interactions whose activities have important consequences for survival, dissemination, and pathogenesis.
Gene bb0562 and adjacent genes bb0563 and bb0564, were all predicted to encode outer membrane proteins containing the conserved domain of unknown function, DUF3996, unique to Borrelia species [9,29], suggesting they may confer similar functions. DUF3996 is also present across two other B. burgdorferi proteins, BB0405 (203 aa) and BB0406 (203 aa) [33, [55][56][57]. The bb0405 and bb0406 genes are co-transcribed, expressed throughout the enzootic cycle and encode immunogenic outer membrane proteins [33,55]. Gene bb0405 is required for B. burgdorferi infection but bb0406 is not [56,57]. The BB0405 homolog in B. afzelii, BAPKO0422, has been shown to be a factor H binding protein, suggesting a role in complement evasion [58]. Northern blot analysis documented primarily monocistronic transcripts for bb0562, bb0563 and bb0564, though longer faint bands were also detected. Given that the RNA was isolated from culture-derived B. burgdorferi, it remains a possibility that some environmental/stress conditions could favor the production of a polycistronic transcript(s). Nevertheless, we did not observe differential expression of any of the three genes throughout the tick-mouse infection model, consistent with other studies [59][60][61][62]. Only gene bb0562, and not bb0563 or bb0564, was found to be important for B. burgdorferi infection and fatty acid-dependent survival, suggesting bb0563 and bb0564 may have disparate roles from bb0562. Furthermore, the lipase function of BB0562 appears to be independent of DUF3996. The catalytic lipase motifs are only present in BB0562 and not the other DUF3996 members. Future work is needed to understand the link, if any, between the five DUF3996 proteins.
Initial characterization of bb0562 began with identification of the contribution of the gene to mouse infection. By both needle inoculation and tick bite transmission bb0562 was important for maximal disseminated infection. Moreover, reintroduction of bb0562 alone to B. burgdorferi lacking bb0562, bb0563 and bb0564 was sufficient to restore wild-type levels of infection. The dissemination defect of B. burgdorferi lacking bb0562 was traced to substantially reduced survival in the blood. The bb0562-dependent infection defect was not ameliorated in immunocompromised mice, suggesting that bb0562 provides a physiological rather than an immune evasion-related function. Combined with the observation of a growth defect of Δbb0562 spirochetes in solid BSK medium, these data led to the hypothesis that bb0562 is important for B. burgdorferi growth and survival in environments deplete in free fatty acids. We found that the rabbit serum component of the in vitro growth media provides the critical source of fatty acids for B. burgdorferi, a fatty acid auxotroph, to scavenge. This growth defect was discovered in solid medium, likely because solid BSK medium is supplemented with 4% rabbit serum, whereas liquid BSKII medium contains 6% rabbit serum, which appears to supply a sufficient amount of free fatty acids for spirochete growth in the absence of bb0562. E. coli lacking the fabA gene, which is essential for biosynthesis of long chain unsaturated fatty acids, is highly attenuated for colony formation on solid media under fatty acid limited conditions [37]. However, this phenotype is less pronounced for growth in liquid media, suggesting that fatty acids are particularly important for bacterial colony formation in vitro [37].
The bb0562-dependent growth defect in low serum media was fully overcome by supplementation with the long chain monounsaturated fatty acid oleic acid. Partial rescue was observed with the addition of the long chain polyunsaturated fatty acid linoleic acid or the long chain saturated fatty acid palmitic acid. These three fatty acids have been shown to be the major fatty acid constituents of the B. burgdorferi membrane when the spirochete is grown in BSK media [19,20,22]. Addition of cholesterol, a non-fatty acid sterol lipid constituent of the B. burgdorferi membrane [21], did not rescue the growth defect of spirochetes lacking bb0562, indicating that the phenotype was fatty acid specific. Strikingly, oleic acid, when provided in the form of a triglyceride, was not able to support the growth of B. burgdorferi lacking bb0562, which led to identification of BB0562 as a novel B. burgdorferi lipase. Indeed, purified, recombinant BB0562 demonstrated lipase activity, which was dependent on the serine residues in the two canonical GXSXG lipase motifs identified in the protein. In bacteria, lipases have been predominantly shown to be secreted extracellularly [40,42] and there are some examples of lipases that localize to the bacterial outer surface [63,64], both of which likely allow interaction of the lipase with target substrates in the environment. Our data indicate that BB0562 is associated with the spirochete outer membrane but is likely not surface exposed. Given this, the mechanism for how the protein interacts with host molecules for fatty acid scavenge remains unknown and warrants further study.
Bacterial lipases are currently classified into 19 families based on amino acid sequence and physiological properties [42]. BB0562 was unrecognized as a putative lipase by conserved domain analysis and rather was annotated as a hypothetical protein of unknown function [29, 65,66]. The findings of our functional analysis of Δbb0562 B. burgdorferi led us to analyze the amino acid sequence of the protein more closely resulting in our identification of two canonical GXSXG lipase motifs and characterization of the lipolytic activity of the protein.
Mutational analysis of the serine residue in each GXSXG motif demonstrated that the enzymatic activity of the protein was dependent on both active sites. Despite the wide diversity of bacterial lipolytic enzymes, to our knowledge CT149, a cholesterol esterase of Chlamydia trachomatis, is the only other protein that harbors two GXSXG motifs [67]. It is interesting to note that serine 34 is conserved across the BB0562 homologs of Lyme disease and Relapsing Fever spirochetes, however, serine 58, while conserved across the BB0562 homologs of Lyme disease spirochetes, is a threonine residue in the homologs of Relapsing Fever spirochetes. It is possible that these active site motif differences result in substrate specificity differences that reflect distinct nutritional needs of the two groups of pathogenic spirochetes. In addition, the apparent higher lipase activities of the recombinant Nus-BB0562 proteins carrying the S58T mutation relative to the other mutant proteins tested, perhaps suggests the ability of a threonine residue in this position to confer some enzymatic activity. BB0562 has a theoretical molecular mass of 19.8 kDa and lacks any cysteine residues, two characteristics of triacylglycerol lipase subfamily I.4, which is comprised of the smallest triacylglycerol lipases primarily PLOS PATHOGENS bb0562 encodes a lipase important for Borrelia infection from Bacillus species [42]. Future work is aimed at biochemical characterization of BB0562 to define the breadth of substrate specificity and function relative to other known lipases.
The lipase activity of BB0562 plays a key role in fatty acid acquisition by B. burgdorferi during infection of the mammalian host, likely via the release of fatty acids from host triglycerides and/or chylomicrons. This model is supported by the significant growth defect of spirochete lacking bb0562 in the blood but not the inoculation site skin tissue. Lipid concentration analysis of mouse tissues and blood has found significantly higher levels of free fatty acids in subcutaneous tissue compared to the amount of free fatty acids in the blood. Moreover, triglyceride concentrations in the blood have been found to be higher than that of free fatty acids [68], suggesting that the blood represents an environment in which the activity of BB0562 may be key for B. burgdorferi survival. No bb0562-dependent growth or survival defect was found for B. burgdorferi in ticks. Free fatty acids are likely released from the blood meal during tick feeding and digestion, bypassing the contribution of the lipase activity of BB0562 for this purpose.
Gene bb0562 is not essential, as loss of bb0562 did not result in absolute lethality of the spirochetes in low fatty acid in vitro or in vivo environments. The survival of B. burgdorferi in the absence of bb0562 may be supported by the lipase activity conferred by other lipases, such as BB0646 [25]. BB0646 harbors a canonical α/β hydrolase fold and a single GXSXG lipase motif [25]. In contrast to BB0562 which appears to be important for scavenge of monounsaturated fatty acids, BB0646 demonstrates specificity for polyunsaturated and saturated fatty acid substrates [25]. In addition, BB0646 exhibits hemolytic activity, which could provide a means for B. burgdorferi to acquire polyunsaturated fatty acids from the phospholipids of host cells rather than triglycerides. Therefore, BB0562 and BB0646 may work in concert to scavenge the full complement of fatty acids required by B. burgdorferi for optimal fitness throughout the enzootic cycle. Although it remains a possibility that the lipase activity of BB0562 also contributes to an immunomodulatory function, as has been demonstrated for other pathogens [69], the finding that Δbb0562 B. burgdorferi remained attenuated in highly immunocompromised mice suggested that BB0562 primarily functions in nutrient scavenge.
In all, this work encompasses the first investigation of gene bb0562 and has led to its identification as a nutritional virulence determinant due to its function as a lipolytic enzyme. Bb-IVET discovered numerous previously uncharacterized genes, many of which are annotated to encode hypothetical proteins unique to Borrelia species. Elucidation of the contributions of non-conserved proteins of unknown function, such as BB0562, to B. burgdorferi survival and infection uncovers new opportunities to gain greater understanding of the unique biology of B. burgdorferi and perhaps discover novel therapeutic targets for the treatment of Lyme disease.

Ethics statement
The University of Central Florida is accredited by the International Association for Assessment and Accreditation of Laboratory Animal Care. Protocols for all animal experiments were prepared according to the guidelines of the National Institutes of Health, reviewed, and approved by the University of Central Florida Institutional Animal Care and Use Committee.

Bacterial strains and growth conditions
Borrelia (Borreliella) burgdorferi clones used in this study were derived from clone B31 A3 and genetic manipulations used infectious low-passage clone A3-68Δbbe02, herein referred to as wild-type (WT) [70]. Cultivation of spirochetes and DH5α E. coli were as previously reported [17].

Northern blot
B. burgdorferi cells grown to 1.15 x 10 8 cells/ml were collected by centrifugation, washed once with 1X PBS (1.54 M NaCl, 10.6 mM KH 2 PO 4 , 56.0 mM Na 2 HPO 4 , pH 7.4) and pellet snap frozen in liquid nitrogen. RNA was isolated using TRIzol (Thermo Fisher Scientific) as described previously [71]. RNA was resuspended in DEPC H 2 O and quantified using a Nano-Drop (Thermo Fisher Scientific).
Agarose northern blots were performed using total wild-type B. burgdorferi RNA as described previously [28,72]. 10 μg of RNA were fractionated on a 2% NuSieve 3:1 agarose (Lonza), 1X MOPS, 2% formaldehyde gel and transferred to a Zeta-Probe GT membrane (Bio-Rad) via capillary action overnight. The RNA was crosslinked to the membranes by UV irradiation and RiboRuler High Range RNA ladders (Thermo Fisher Scientific) were marked by UV-shadowing. Membranes were blocked in ULTRAhyb-Oligo Hybridization Buffer (Ambion) and hybridized with 5 0 32 P-end labeled oligonucleotides probes (S1 Table). After an overnight incubation, the membranes were rinsed with 2X SSC/0.1% SDS and 0.2X SSC/0.1% SDS prior to exposure on film. Blots were stripped by two 7-min incubations in boiling 0.2% SDS followed by two 7-min incubations in boiling water.

B. burgdorferi genetic manipulations
Targeted gene deletion and cis complementation was carried out via allelic exchange on the B. burgdorferi chromosome. All oligonucleotides used are listed in S1 Table. Spectinomycin/ streptomycin (flaBp-aadA) and gentamycin (flgBp-aacC1) resistant cassettes were Phusion-PCR amplified from B. burgdorferi shuttle vectors [15,73,74]. Allelic exchange constructs were engineered via a PCR-based overlap extension strategy [27] using Phusion-PCR and target specific oligonucleotides and PCR-generated linear DNA transformed into the appropriate B. burgdorferi clones, as described [75]. All clones were verified by PCR using oligonucleotides external to the lesion (S1 Table and S1 Fig). All clones were verified to contain the endogenous plasmid content of the parent [76,77].
Gene bb0562 along with 145 bps of upstream sequence were amplified with and without a 3Xflag epitope sequence [78] at the C-terminus via Phusion-PCR using B. burgdorferi A3 genomic DNA (S1 Table) and cloned into pBSV2G [74] in E. coli. The plasmids were verified by PCR, restriction enzyme digest, and Sanger sequencing, prior to transformation into Δbb0562 and/or Δbb0562-bb0564 B. burgdorferi, as described [75].

Growth analysis
Growth curves were performed in biological triplicate over up to 168 hours as described [17].

RNA extractions and RT-qPCR analysis
Isolation of RNA from in vitro grown log phase (3-7x10 7 cells/ml) B. burgdorferi was performed using TRIzol reagent (Life Technologies). For isolation of RNA from infected ticks, C3H/HeN mice (Envigo) were needle inoculated with B. burgdorferi B31 A3, at a dose of 10 5 spirochetes and infection confirmed via positive seroreactivity to B. burgdorferi lysate, as previously described [77,88]. Three weeks post inoculation, approximately 200 naïve~5 month old Ixodes scapularis (Centers for Disease Control, BEI resources) per mouse were fed to repletion as previously described [89,90]. Approximately one week following feeding to repletion, triplicate groups of 25-50 fed larvae and 13-15 fed nymphs, or five weeks following feeding to repletion, triplicate groups of 25 unfed nymphs, were snap frozen in liquid nitrogen. Ticks were homogenized in~1 ml TE buffer (10 mM Tris, 1 mM EDTA, pH 8.0) containing 0.5 mg/ ml lysozyme in gentleMACS TM M tubes using a gentleMACS TM Dissociator, setting RNA 2 (Miltenyi Biotech). For RNA isolation from infected mouse tissue, mice were inoculated as described and 10 days post-inoculation the animals were sacrificed, and bladder tissues harvested. In triplicate groups of 3 bladders per RNA preparation, bladders were snap frozen in liquid nitrogen and homogenized as described for infected ticks. Mouse and tick samples were transferred to 1.5 ml safe-lock Eppendorf tubes, with 1% SDS and incubated at 64˚C for 2 min, followed by the addition of 0.1 M NaAOc pH 5.2. RNA extraction continued using a hot phenol protocol as described previously [28,91].
Up to 50 μg of RNA was treated with 10 U DNase I (Roche) and 80 U rRNasin (Promega) at 37˚C for 15 min. The RNA samples were purified as described [91]. The DNase treatment and purification were repeated twice to ensure DNA removal from samples.

Mouse infection by needle inoculation
All B. burgdorferi clones were grown to stationary phase (~10 8 /ml) and diluted to the desired inoculum density. Six-to-eight-week-old female C3H/HeN mice were purchased (Envigo) and six-week-old male NOD-scid IL2Rgamma null (NSG) mice were bred at the UCF vivarium from purchased breeding pairs (The Jackson Laboratory). Groups of 6 mice each were needle inoculated with 10 4 B. burgdorferi intradermally (C3H/HeN) or with the dose split 4:1 between the intraperitoneal and subcutaneous routes (NSG). Inoculum densities were confirmed by plating for individuals in solid medium or by limiting dilution in liquid medium in 96-well plates. All inoculum cultures were verified to contain the expected endogenous plasmids [76,77] and individuals from each inoculum clone were analyzed for the presence of virulence plasmids lp25, lp28-1 and lp36 [89]. Mouse infection was determined by spirochete reisolation from tissues and quantitative PCR for B. burgdorferi loads in tissues [35,77].

Determination of B. burgdorferi density in blood
On day 6 post-inoculation approximately 50 μl of blood was collected from the submandibular vein of each mouse into K2 EDTA MiniCollect blood collection system tubes (Greiner). For each blood sample, two volumes of blood, approximately 38 μl and 4.8 μl, were each added to 40 ml of liquid BSKII containing RPA cocktail (60 μM rifampicin, 110 μM phosphomycin, and 2.7 μM amphotericin B) and plated for colony forming units (CFU) across two 96-well plates. Plates were incubated at 35˚C with 2.5% CO 2 for 14 days and scored for the presence or absence of spirochete growth. Each positive well was considered a single CFU. B. burgdorferi/ ml of blood was calculated as the number of CFUs/ volume of blood analyzed multiplied by 100 for each sample and the average of the two measurements for each sample was determined.

Infection of Ixodes larvae by immersion and tick feeding
Approximately 4-month-old Ixodes scapularis naïve larval ticks (CDC, BEI resources or Oklahoma State University) were dehydrated by exposure to saturated ammonium sulfate for 24-48 hr. Log phase grown B. burgdorferi clones were diluted to 2 x 10 7 cells/ml in BSKII. 500 μl of spirochetes were incubated with dehydrated ticks at 35˚C for~1.5 hrs and washed twice with PBS [92]. The inoculum cultures were verified to contain the expected endogenous plasmids [76,77]. Infectious plasmids lp25, lp28-1 and lp36 were present in 80-100% of individuals of each inoculum culture. Three days post artificial infection cohorts of~150 larvae were fed to repletion on groups of 3-7 naïve C3H/HeN mice (Envigo) per B. burgdorferi clone. Subset of fed larvae were allowed to molt into nymphs and subsequently cohorts of 25 nymphs fed on groups of 3-7 naïve C3H/HeN mice (Envigo) per B. burgdorferi clone. Three weeks post nymph feeding, mice were assayed for infection by spirochete reisolation from tissues [35,77].

PLOS PATHOGENS
bb0562 encodes a lipase important for Borrelia infection

Determination of B. burgdorferi densities in ticks
The B. burgdorferi densities in ticks were assessed before and after feeding to repletion on mice by performing qPCR analysis of total DNA isolated from ticks or plating of dilutions of whole triturated ticks in solid or liquid medium [77,89,93]. Total DNA was isolated from groups of 17-50 unfed larvae, 5-13 fed larvae 7-14 days post feeding, and 6-10 unfed nymphs. Ticks were homogenized with a plastic pestle in a microfuge tube, and genomic DNA was isolated using a NucleoSpin tissue kit (Clontech Laboratories) according to the manufacturer's specifications. Individual fed nymphs were surface sterilized and assessed for colony forming units in solid or liquid medium containing RPA cocktail [89,90].

Low serum/fatty acid growth analysis
For initial assessment of B. burgdorferi growth in solid medium, in biological triplicate, B. burgdorferi clones were inoculated into complete BSKII liquid medium at a starting density of 1x10 5

Cloning and purification of recombinant BB0562
The wild-type bb0562 gene was PCR amplified using Q5 polymerase (New England Biolabs) and primers 2643 + 2644 (S1 Table). The serine 34 and/or serine 58 bb0562 point mutants were generated using overlapping PCR (S1 Table). The PCR fragments were cloned into pET43.1b in E. coli, generating 6Xhis, nus N-terminal and 6Xhis C-terminal fusion constructs. All fusion constructs were verified by restriction digest and Sanger sequencing. The pET43.1b-bb0562 plasmids were moved to NiCo21 (DE3) E. coli (New England Biolabs) for protein induction and purification. The E. coli clones were grown to late logarithmic phase in lysogeny broth shaking at 37˚C, chilled at 4˚C for 1 hour and induced with 0.5 mM IPTG at 18˚C for 16 hours. Cells were harvested by centrifugation at 4˚C, resuspended in 10 ml of His-binding buffer (Cytiva) and disrupted by sonication. Total protein lysates were cleared of insoluble debris by centrifugation at 4˚C. The soluble lysate was loaded by gravity flow onto nickel beads and washed according to the manufacturer's instructions (Cytiva). Protein was eluted from the nickel beads in a stepwise fashion using 100 mM to 500 mM imidazole. The 300 mM imidazole fractions were selected for all downstream applications (S6 Fig). The protein samples were moved into PBS using Zeba spin desalting columns (Thermo Fisher Scientific). The proteins were quantified by BCA protein assay (Pierce) and stored at 4˚C.

PLOS PATHOGENS
bb0562 encodes a lipase important for Borrelia infection

Lipase activity assay
Lipase activity was determined using the Lipase detection kit III (fluorometric) containing a proprietary long chain fatty acid triglyceride derivative substrate (Abcam). Two micrograms of each recombinant protein, including Nus alone, were used. The assay was performed in triplicate according to the manufacturer's instruction (Abcam) over a 300-minute time course. The amount in picomoles (pmol) of fluorescent methylresorufin produced by hydrolysis of the lipase substrate by each recombinant protein at each time point was extrapolated from a standard curve. Lipase activity was calculated as pmol/min/μg.  Fig. Gene bb0562 is sufficient to restore wild-type levels of spirochete load in tissues to Δbb0562-bb0564 B. burgdorferi. Groups of 6 immunocompetent C3H/HeN mice were needle inoculated intradermally with 10 4 B. burgdorferi per clone. Three weeks post-inoculation ear, heart and joint tissues were collected, and total DNA was extracted. B. burgdorferi load was measured by quantifying B. burgdorferi flaB copies normalized to 10 6 mouse nid copies using quantitative PCR. Each data point represents an individual mouse. The mean is indicated by a horizontal black line. Statistical analyses were performed using the unpaired t-test, GraphPad